Ready to make an impact

.

AI/ML Researcher & Engineer.

Working on AI evaluations

I build machine learning systems and research how to evaluate them and make them safer.

Selected work
Résumé
Dhruv Tibarewal
Chennai · IN

Currently

01

M.S. by Research

IIT Madras

Data Science & AI

02

AI Safety

Secure AI Futures Lab

Evals & red teaming

03

Developer

ML Hub

AI features & fixes

04

Top-Rated

Upwork

AI/ML Developer

01

Research and
engineering.

The three areas my work sits in.

  1. 01

    Trustworthy AI

    Privacy, evaluation, and harmful-response behaviour, with a focus on multilingual and low-resource settings.

  2. 02

    Production ML

    End-to-end ML systems, from data ingestion to deployed and monitored inference.

  3. 03

    Agentic Systems

    Retrieval and multi-agent workflows, built with evaluation and guardrails.

02Live

See how it connects.

The domains I work in, the organisations I have worked with, and the projects that came out of them.

30 nodes · 48 edges

  • Domain
  • Organisation
  • Project

Hover to trace · click to select · drag to move

30n / 48e

Force-resolved
03Selected work

Open any case study for the technical approach, scope, and outcomes, not just a headline.

ControlPlane · The Tower interface

Live site: the value-of-information cascade, layer by layer

Hover to play
01 / 04
01Accenture Innovation Challenge 2026 Live

Real-time oversight for enterprise AI, run as one economic decision. A drop-in gateway that judges every response on performance, cost, and responsibility, and buys a check only when it could change the outcome.

HaluEval: F1 0.80 vs 0.76 at equal recall · ~55% fewer expensive checks

  • Value-of-Information
  • Conformal risk control
  • HHEM-2.1
  • FastAPI
  • Next.js
  • OpenAI-compatible API
Live demo Repository
Chimera interface

Live site: the hardening curve closing round by round

Hover to play
02 / 04
02Mastercard Innovation Challenge 2026 Live

A closed-loop adversarial AI lab for GenAI-era payment fraud. An AI red team evolves fraud against a live detector; the detector retrains on what gets through. Identify, generate, defend as one loop.

Red team drops recall 72.9% → 19.2% · retraining restores 82.9%

  • LangGraph
  • LightGBM
  • GraphSAGE GNN
  • Isolation forest
  • FastAPI
  • Next.js
Live demo Writeup Repository
Aegis · Multi-Agent Erasure Verification interface

Live site: the four-layer erasure audit

Hover to play
03 / 04
03AI safety · Privacy research Live

An evaluation harness that tests whether machine "erasure" is real across four layers of an ML stack: the database, the model weights, the vector index, and the semantic cache.

Two independent unlearning engines · one shared rubric

  • LangGraph
  • LoRA
  • LiRA
  • Task Vector Negation
  • Qdrant
  • HF Spaces
Live demo Repository
Counterpoint Engine interface

Live site: four agents, one cyclic LangGraph workflow

Hover to play
04 / 04
04Agentic AI · Autonomous research Live

A cyclic multi-agent research system. Give it a claim; four agents argue it out live, re-searching themselves until confidence clears 80% or they run out of cycles.

Live WebSocket streaming · nothing precomputed

  • LangGraph
  • FastAPI
  • Celery
  • Playwright
  • Neo4j
  • WebSockets
Live demo Repository
Further work10 entries
  • 05

    Indic Multimodal Trust & Safety

    Ongoing research

  • 06

    Renewable Forecasting at Scale

    Production ML

  • 07

    Physics-Guided ML for Catalysis

    ML Hackathon · IIT Kharagpur 2026

  • 08

    Machine Unlearning for Recommenders

    B.Tech thesis

  • 09

    Data Intelligence Graph Engine

    Production data systems

  • 10

    AIInterviewCoach

    BlockVerse · HR AI product

  • 11

    Insight-Hire

    BlockVerse · HR analytics

  • 12

    Responsible Mental Health Platform

    Open-source collaboration

  • 13

    Web Query Agent

    Agentic AI · Personal project

  • 14

    LLMs Analytical Reasoning

    Hackathon · ML reasoning

04Experience

A timeline of roles where I have helped turn AI ideas into research, systems, and outcomes.

Current
01Jul 2026 - Present

Special Research Fellow · Evals & Red Teaming

Secure AI Futures Lab
/New Delhi · Hybrid · Part-time
  • Contribute to AI-safety research on hallucinations and the evaluation of Indic open-source models.
  • Maintain India state-wise AI policy data and state policy documentation to support research on AI safety in India.
  • Work on the India AI Governance Monitor, a sourced, interactive record of India's AI ecosystem built for researchers, students, and policymakers.
  • Expanding the RFP-matching database to ~4,000 profiles by adding faculty from top Asian universities.
AI safetyLLM evalsRed teamingAI policyIndic languages
Read more
02

Developer Volunteer

ML Hub
/Remote
  • Helping the team bug-fix critical issues before their product launch.
  • Implementing AI features and integrating intelligent capabilities into the platform.
Software engineeringAI featuresBug fixingPre-launch
Read more
03

Top-Rated AI/ML Developer

Upwork
/Remote
  • Deliver end-to-end AI products spanning RAG chatbots, time-series systems, semantic search, and agentic automation workflows for global clients.
  • Built an HVAC support RAG system and business-process automations using LangGraph, MCP, Zapier, and MindsDB for a Florida-based company.
  • Developed a vector-embedding SEO/GEO proof of concept across 33 websites, from scraping to content-gap scoring.
  • Maintain a Top-Rated profile with consistent positive client feedback across diverse AI/ML engagements.
RAGLangGraphAutomationClient deliverySEO/GEO
Read more
Previously
04Mar 2026 - Jun 2026

Technical Operations Intern

SteadRise
/New Delhi · Hybrid
  • Built production-ready scripts for web-scale data sourcing, scraping, validation, enrichment, and provenance-aware profile consolidation.
  • Built a graph exploration engine for dense professional-network data: 10k+ nodes, 5k connections, layered views, and path-finding.
  • Built the professor/RFP matching workflow: scraped CS faculty across 80+ CFTIs and scored ~1,950 profiles with keyword filters, embeddings, and an LLM judge.
  • Contributed to research on AI agents, red teaming, AI-safety organisations, and governance.
Data systemsKnowledge graphsRFP matchingPythonAI governance
Read more
05Feb 2026

AI Consultant

Jaagruk Bharat
/New Delhi · Remote
  • Diagnosed and helped fix an underperforming schema-finder workflow serving 50,000+ Indian government schemes.
  • Improved retrieval quality and accuracy for a large civic-tech corpus targeting scheme accessibility.
Civic techInformation retrievalSchema matching
Read more
06Jul 2025 - Dec 2025

Machine Learning Engineer · Renewable Forecasting

NewGrid Consulting LLC/Arizona, US · Remote
  • Led an end-to-end forecasting platform for solar and wind operations: ingestion, feature engineering, training, inference, monitoring, and reporting.
  • Productionized LightGBM, CatBoost, and XGBoost ensembles after evaluating ARIMA and LSTM approaches; used 30-40 weather and operational features.
  • Deployed 1,700+ packaged models across ~450 viable plants with 98% deployment success; internal production achieved average R² ≈ 0.73 and RMSE ≈ 0.86 MW.
  • Implemented weekly automated inference with FastAPI and cron workflows, UTC-safe time-series handling, and quantile-regression prediction intervals.
ForecastingMLOpsFastAPILightGBM
Read more
07Jun 2025 - Nov 2025

AI Researcher

CSIRO’s Data61
/Australia · Remote
  • Researched privacy threats and defenses in federated learning with Australia’s national science agency.
  • Explored implementation of data-reconstruction attacks to understand privacy leakage in distributed ML environments.
  • Studied various attack and defense surfaces in FL, including gradient leakage, model inversion, and defense mechanisms.
  • Worked at the intersection of machine learning, privacy protection, distributed systems, and secure system design.
Federated learningPrivacyAI safetyDistributed systems
Read more
08Feb 2024 - Jul 2024

AI Intern

BlockVerse Institute
/Netherlands · Remote
  • Completed a 50-hour applied AI/ML cohort covering GenAI, LLMs, supervised ML, NLP, computer vision, and MLOps.
  • Built end-to-end HR-facing AI products including AIInterviewCoach and Insight-Hire.
  • Delivered nearly 10 hands-on projects across classroom and product-oriented ML use cases.
GenAINLPComputer visionMLOps
Read more
09Feb 2024 - Apr 2024

Junior ML Engineer

Omdena
/Remote
  • Led RAG development and Llama Guard integration for an open-source global mental-health assistance chatbot.
  • Collaborated with 10+ contributors and explored DPO-based optimization for responsible LLM assistance.
RAGLlama GuardDPOOpen source
Read more
05Education

Degrees and current research programme.

Jun 2026 - Jun 2028
01

M.S. by Research · Data Science & AI

Indian Institute of Technology Madras

Researching multimodal deepfake detection for Indic languages under the India AI Mission, guided by Dr. Krishna Pilutla and Prof. Balaraman Ravindran.

AI researchMultimodal AIMulti-agent systemsRAGDCC Batch Representative
Read more
Oct 2022 - May 2026
02

B.Tech · Computer Science (AI & ML)

Heritage Institute of Technology

Graduated with a 9.22/10 GPA while building foundations across AI/ML, data systems, deep learning, and DevOps.

GPA 9.22/10AI/MLDevOpsData systems
Read more
06Technical toolkit

Grouped by the kind of work they are used for.

4 groups / 40 entries

AI, Research & Evals

01
  • 01LLMs & Generative AI
  • 02Multi-agent systems
  • 03LangChain · LangGraph
  • 04RAG architecture
  • 05AI safety & red teaming
  • 06LLM evaluations
  • 07Federated learning
  • 08Machine unlearning
  • 09NLP & transformers
  • 10PyTorch · TensorFlow

Production ML & MLOps

02
  • 01Time-series forecasting
  • 02LightGBM · CatBoost · XGBoost
  • 03MLflow & model monitoring
  • 04Docker · Kubernetes
  • 05FastAPI
  • 06CI/CD · Jenkins · Ansible
  • 07AWS · GCP · Azure
  • 08Feature engineering
  • 09Model validation
  • 10Quantile regression

Data & Engineering

03
  • 01Python · SQL
  • 02Pandas · NumPy · scikit-learn
  • 03Vector databases · ChromaDB
  • 04OpenAI · Hugging Face
  • 05Playwright · Firecrawl
  • 06MongoDB Atlas
  • 07REST APIs
  • 08Git · GitHub
  • 09Plotly · Power BI
  • 10Streamlit

Research & Data Practice

04
  • 01Experimental design
  • 02A/B testing
  • 03Statistical analysis
  • 04EDA
  • 05Data quality review
  • 06Entity resolution
  • 07Web scraping
  • 08Schema design
  • 09Technical documentation
  • 10Stakeholder communication
07Kaggle

Competition practice.

Hands-on tabular ML on Kaggle. Eleven competitions entered; these are the strongest finishes.

Visit Kaggle profile
CompetitionField
01

Multi-Class Prediction of Obesity Risk

Playground S4E2 · multiclass classification

Ranked 84 / 3,587

02

mlX Stage 1 Regression: Predict my Marks!

Community · regression

Ranked 3 / 44

03

Regression with an Abalone Dataset

Playground S4E4 · regression

Ranked 177 / 2,606

04

Predicting Smartphone Addiction

Playground S6E8 · 2026

Ranked 1,204 / 3,531

05

Binary Classification of Insurance Cross Selling

Playground S4E7 · binary classification

Ranked 866 / 2,234

08Certifications

Selected certifications across AI, ML, cloud, and software engineering.

16 entries · hover to pause

Building Multi-Agent Systems

CrewAI

Jul 2024
Multi-agent SystemsLLMsAgent memory

Data Sage Hackathon

IIT Gandhinagar

Jul 2025
Data Science

ML Hackathon

IIT Gandhinagar

Jul 2025
Machine Learning

Function Calling and Data Extraction with LLMs

DeepLearning.AI

Jun 2024
Function CallingLLMs

Prompt Engineering with Llama 2 & 3

DeepLearning.AI

Apr 2024
Prompt EngineeringLlama 2

Machine Learning

Stanford University / DeepLearning.AI

Jan 2024
Machine Learning

Neural Networks and Deep Learning

DeepLearning.AI

Jan 2024
BackpropagationDeep Learning

Generative AI

Google Cloud

Jan 2024
GCPGenerative AI

Large Language Models

Google Cloud

Jan 2024
GCPLLMs

Responsible AI

Google Cloud

Jan 2024
GCPResponsible AI

Machine Learning on AWS

Amazon Web Services

Jan 2024
MLAWS

Feature Engineering

Kaggle

Feb 2024
Feature EngineeringML

SQL for Data Science

UC Davis

Feb 2024
SQLData Analysis

Intermediate Machine Learning

Kaggle

Sep 2023
PipelinesData Leakage

100 Days of Code: Python Pro Bootcamp

Udemy

Jan 2024
PythonWeb Development

Mastering DSA using C and C++

Udemy

Nov 2023
C++Algorithms
09Writing

I am putting together writing on AI safety, agentic systems, and lessons from shipping ML. First posts are on the way.

In progress
  • 01AI safety
  • 02Agentic systems
  • 03Shipping ML in the real world
Notify me when the first post drops
10Contact


I am interested in research collaborations, thoughtful ML products, and teams working on reliable, safe AI. Let's talk.

dhruvtibarewal [at] gmail [dot] com
LinkedIn
GitHub